An ML engineer is using AWS Glue to transform proprietary data from a third-party vendor to a format that the ML engineer intends to use with the Amazon SageMaker DeepAR forecasting algorithm. The dat

Question

Options

• AUse Snappy to convert the files to RecordIO-Protobuf and to compress the files.
• BUse XZ to convert the files to RecordIO-Protobuf and to compress the files.
• CUse XZ to convert the files to Apache Parquet format and to compress the files.
• DUse gzip to convert the files to Apache Parquet and to compress the files.

• A
  2% (1)
• B
  12% (7)
• C
  81% (46)
• D
  5% (3)

Explanation

Amazon SageMaker DeepAR natively supports two input formats: JSON Lines and Apache Parquet. RecordIO-Protobuf (referenced in options A and B) is the native format for many other SageMaker built-in algorithms (e.g., Linear Learner, XGBoost in pipe mode) but is NOT supported by DeepAR-eliminating options A and B. Between the remaining options, XZ compression (option C) provides a significantly better compression ratio than gzip (option D), which is important for optimizing storage costs on a large multi-file time series dataset. Apache Parquet is also a columnar, efficiently encoded format that further reduces storage footprint. Therefore, converting to Apache Parquet and compressing with XZ provides the best compatibility with DeepAR and the best storage cost optimization.

No community discussion yet for this question.